Abstract
Sensitivity experiments with sawtooth gratings show that all spectral components of the sawtooth interact. Interpreting the sensitivity function as the transfer function of the system, a formula can be derived by which the detection threshold of arbitrary patterns can be predicted, if the shape of their spectrum is known. This formula is tested with a variety of test patterns. For certain composite patterns this prediction fails. In these cases it can be shown that the threshold is determined by the stronger component. The question whether the postulated matched filter really exists, or whether the experimentally obtained performance is the result of some nonlinear behavior is discussed.
Original language | English |
---|---|
Pages (from-to) | 59-64 |
Number of pages | 6 |
Journal | Photogr Sci Eng |
Volume | 22 |
Issue number | 2 |
State | Published - 1978 |